MULTIVARIATE NON-NORMAL DISTRIBUTIONS AND MODELS OF DEPENDENCY
MULTIVARIATE NON-NORMAL DISTRIBUTIONS AND MODELS OF DEPENDENCY
复制标题
多元非正态分布和依赖性模型
DOI：
--
复制
发表时间：
1994
期刊：
影响因子：
--
通讯作者：
I. Olkin
中科院分区：
文献类型：
--
作者：
I. Olkin
单变量和多元正态分布在统计建模中发挥着核心作用。然而，有许多自然现象并不符合正常规律。特别是，需要对多元二项分布、泊松分布、指数分布、伽马分布和贝塔分布进行建模。有很多方法可以创建具有给定边际的双变量（或多变量）分布，因此了解这些扩展的基本比率非常重要。本综述概述了许多用于创建“自然”多元非正态分布的方法。
The univariate and multivariate normal distributions play a central role in statistical modeling. However, there are many natural phenomena that do not behave according to the normal law. In particular, there is a need to model multivariate binomial, Poisson, exponential, gamma, and beta distributions, for example. There are many ways to create bivariate (or multivariate) distributions with given marginals, so that it is important to understand the underlying ratio-nale for these extensions. The present review outlines a number of methods that have been used to create "natural" multivariate non-normal distributions.